Nitrogen oxides (NOx ≡ NO + NO2) play a central role in air pollution. Super-emitters present unique opportunities for emission mitigation in China and beyond. They comprise intensive industrial facilities (e.g., power or chemical plants), less than 1 × 1 km2 with high NOx plumes, dominating localized concentrations within a limited geographical scope. However, identification of super-emitters typically challenges emission mitigation due to insufficient spatiotemporal resolution. Here we map NOx emissions using an efficient, super-resolution (1 × 1 km2) inverse model based on whole-year TROPOMI satellite observations. Our map offers unique insights on nationwide NOx super-emitters. We resolve 1625 super-emitters in virtually every corner in China, even in remote and mountainous zones, which we trace back to either an industrial hotspot or a cluster (i.e., an industrial park). A state-of-the-art bottom-up emission inventory MEICv1.3 largely (67%) agrees with our results within a factor of two for cities. However, that inventory does not identify these super-emitters, particularly underestimating one-third of the emissions by at least an order of magnitude. Even, nearly up to three-quarter super-emitters are found to be displaced or missing. Moreover, traditional top-down inverse methods do not effectively detect such super-emitters. Here we show it is necessary to address the NOx budget by revisiting super-emitters on a large scale. Integrating the results we obtain here with a multi-tiered observation system can lead to identification and mitigation of anomalous NOx emissions.